Mindfulness-Based Stress Reduction (MBSR), an 8-week group psychosocial intervention, has received increasing research support for its efficacy in oncology settings. Changes associated with MBSR participation for people with cancer include improved psychological functioning and quality of life. However, as with other populations, it remains unclear which components of MBSR bring about change and whether targeted constructs are critical in changing outcomes. We propose a mediation model to be tested as a first step towards understanding program mechanisms. Specifically, changes in mindfulness and rumination were hypothesized to mediate the impact of MBSR participation on symptoms of depression in people living with cancer. A waitlist-controlled study of MBSR participation in 77 women who had completed cancer treatment was conducted to test this model. Pre- to post-program, MBSR participants improved significantly more on depressive symptoms and mindfulness and decreased more on rumination scores compared to waiting controls. Decreases in rumination mediated the impact of MBSR on depressive symptoms, but mindfulness scores did not. Methodological recommendations are presented to promote research that will further elucidate the mechanisms of action of MBSR. Mediation analyses will inform the next generation of randomized controlled trials and may lead to program modifications that will maximize the effectiveness and efficiency of mindfulness-based interventions in cancer settings.